Re: RV: na'e entails na?

[HACKER G N <c9709244@ALINGA.NEWCASTLE.EDU.AU>]

On Sat, 20 Sep 1997, JORGE JOAQUIN LLAMBIAS wrote:

> Ok, once again I have been persuaded by And to change my
> mind. My position now is that na'e by itself does not entail na.
> It only does so when the selbri in question partitions its domain
> into exclusive regions (I try to explain what I mean by this below).
>
> >For example, everyone is either citizen of France or citizen of
> >some other country. [NB INCLUSIVE OR] I want to describe
> >the latter group as "na`e fraso zei selgugde"
> [...]
> > but will not be
> >able to if everyone bar me gets their way!
>
> I now agree with your position, as long as it is clear that {na'e broda}
> asserts not just any relationship other than broda. It must claim that a
> relationship from a very reduced group holds among the arguments.
> For the case of fraso, the relationships that may hold can be glico,
> dotco, spano, brito, etc, but not for example ropno, since {ko'a ropno}
> does not allow us to conclude that {ko'a na'e fraso}. In the case of
> glico we cannot have brito as one of the possible "others", and so on.

That's fair enough, because it is consistent with the original intention
of "na'e" to imply a scale. In this case, the scale must be sensibly
construed so that "na'e" actually says something interesting. However,
this interesting statement need not in itself entail "na" -  which is
itself an interesting conclusion!

So subject to that proviso, I agree with you too, And. There, now you
shouldn't feel so picked on! ;)

 > > How this very restricted group of
> relationships is selected is
> the > difficult part, and probably very context dependent.

I agree, which is probably why the relationship between "na'e" and "na"
wasn't covered in clear detail in the refgramm.

>In many cases
> the domain of arguments gets partitioned into exclusive regions
> by the predicates, and then na'e does entail na. For example,
> taking {zmana'u} to mean "x1 is positive", then {ko'a na'e zmana'u},
> "k is non-positive", does entail {ko'a na zmana'u}, because the only
> possibilities left are that k is negative or that k is zero. All other
> relationships that may be true of ko'a are irrelevant.

Yes, and moreover I think you can consider this case a subset of the
abovementioned cases. It simply falls at one end of extremity along a
continuum of restricted relations that make meaningful claims that pertain
to a relation with a scalar negation.

 >
> With this strong restriction, I think there isn't really that much of
> a distance between the strong and weak forms of na'e. In many
> cases it makes no difference which one we choose. I prefer the
> weak form because, as And pointed out, the strong form can be
> easily obtained with an end-of-bridi naku, whereas the weak
> form cannot.

Except that I think the issue is made obsolete by the fact that both
restrictions can be properly related to each other along the continuum
described above, a third unifying view, which has the positive claim
associated with "na'e" logically related to the scalar-negated thing in a
way that makes a meaningful claim - context-dependent and with "na" only
being entailed in the most restrictive of scales, AS HAPPENS when the
choice between "da" and "na'e da" is exclusive, such as with positive
numbers and na'e positive numbers.

Geoff